Start with the problem, not a product.
Every job is different. We first understand what is taking time, causing errors or making the company harder to run. I then give you a plan and roadmap. Only after that do we decide what, if anything, should be built and how you want to pay for it.
How working together starts
See the quality of the thinking before deciding whether to go further.
Free initial consultancy
We talk to the people who understand the work day to day, explore the problems that have frustrated them and clarify what the business would like to achieve. You do not need a specification or know whether the answer involves AI.
A plan and roadmap
I turn what we learn into practical options: the desired outcome, the sensible order of work, likely costs and risks, and what success should look like. The answer may be process improvement, integration, automation, bespoke software or no build at all.
A commercial model that fits
If you want me to deliver the plan, the build can sometimes be absorbed into a monthly licence and support subscription. You can also buy the software outright and choose a support package. Everything is scoped and agreed before work starts.
The solution should be proportionate to the problem
Not every awkward workflow needs new software. Sometimes a process can be simplified. Sometimes two existing systems need connecting, or one repetitive step can be automated. Those are often the quickest and least disruptive wins.
Where focused or bespoke software is the right answer, modern development tools have changed the economics. Small improvements can be delivered quickly, and larger builds can be phased around the business rather than attempted as a risky big-bang project.
Bespoke software used to be reserved for large corporations spending hundreds of thousands of pounds. AI-assisted development has reduced the time and cost dramatically, making software shaped around a particular business realistic for far more companies.
More than one way to work together
The commercial arrangement should fit the solution and the way you prefer to buy.
Licence and support
Where the project suits it, some or all of the build cost can be absorbed into a monthly subscription covering the software licence, hosting and ongoing support.
Buy the software
You can purchase the software outright. I can then provide the level of ongoing support, maintenance and improvement that your team needs.
Strengthen what you build
If your team is developing its own AI systems, I can help make them more robust, controlled and supportable, then provide ongoing cover so you can focus on the business.
Before you spend anything
The conversation is free. It's the build you pay for.
Most jobs start with a conversation, not a proposal. Tell me about the workflow eating your team's week, the workaround everyone relies on, or the idea you haven't had time to test — and I'll tell you what's possible, what isn't, and what I'd do. If replacing your system is the wrong move, I'll say that too.
You bring the business. I bring the build.
You don't need to arrive with a specification. The people running a business shouldn't also have to be system architects. I've seen too many projects where a developer builds exactly what was asked for, the result doesn't solve the problem, and the customer gets blamed for writing the wrong spec. You describe the outcome; working out the right shape for it is my job.
If it's small, I'll build it free so you can see it working.
Where a demonstration is small enough to do in a couple of days, I'll build it using one anonymised or sample job at no charge. It is a way to test the idea, not production software: live integrations, deployment and ongoing support are scoped separately. There's no obligation afterwards.
Start small, go for the low-hanging fruit.
The biggest mistake I see is trying to plan everything up front. It's the approach hourly consultancies favour — the longer the scoping, the bigger the invoice — but you get much better results by picking one focused problem, shipping it, and going from there. Confidence builds with each result, and you quickly learn where the next bit of leverage is. Plenty of jobs start with a single improvement alongside the existing system. Replacement is considered only if the evidence eventually supports it.
Start a conversation — no chargeIs one workflow taking more time than it should?
Talk me through what happens now. I'll help you work out whether there is a practical improvement worth making.